J. Phys. Chem. B 2001, 105, 6211-6220
6211
Supported Tantalum Oxide and Supported Vanadia-tantala Mixed Oxides: Structural Characterization and Surface Properties Michael Baltes,*,† Arla Kyto1 kivi,‡ Bert M. Weckhuysen,§ Robert A. Schoonheydt,| Pascal Van Der Voort,† and Etienne F. Vansant† Laboratory of Adsorption and Catalysis, Department of Chemistry, UniVersity of Antwerpen (U.I.A.), UniVersiteitsplein 1, 2610 Wilrijk, Belgium, Fortum Oil and Gas Oy, Oil, Research and DeVelopment, POB 110, 00048 Fortum, Keilaranta 6, Espoo, Finland, Department of Inorganic Chemistry and Catalysis, Debye Institute, UniVersity of Utrecht, Sorbonnelaan 16, PO Box 80083, 2508 TB Utrecht, The Netherlands, and Centrum Voor OpperVlaktechemie en Katalyse, Departement Interfasechemie, K.U. LeuVen, Kardinaal Mercierlaan 92, 3001 HeVerlee, Belgium ReceiVed: February 16, 2001; In Final Form: April 20, 2001
Silica-, alumina-, and zirconia-supported tantalum oxide and supported vanadia-tantala mixed oxide catalysts were investigated by XRD, UV-vis-DR, FTIR, and FT-Raman spectroscopy to determine the nature of the surface tantala species. Two types of surface TaOx species were identified. On the SiO2 support, at low coverages tetrahedral TaO4 species exist, with a TadO bond and three bridging Ta-O-support bonds. At higher coverages, octahedral TaO6 species occur. On Al2O3 and ZrO2 both surface species exist already at considerably low Ta loading. The acidic properties of the catalysts were investigated by an FTIR study of adsorbed pyridine. Supported tantala is clearly more Lewis acidic in comparison to vanadia. The oxidation of methanol was used to probe the catalytic performance of the supported oxide species and from the activity and product distribution it is concluded that supported tantala predominantly shows acidic over redox behavior.
Introduction Group V transition metal oxides are very efficient as catalysts in numerous processes. The physical, chemical and catalytic properties of supported vanadium and niobium oxides have been studied extensively in the past, using several spectroscopic characterization techniques and catalytic reactions. Vanadium oxide catalysts catalyze a number of industrially important processes. They have been used successfully in the selective oxidation of aliphatic and aromatic hydrocarbons (e.g., oxidation of methane to formaldehyde,1 oxidation of o-xylene to phthalic anhydride2-4), the selective oxidation of SO2 in the production of sulfuric acid,5 and the selective catalytic reduction (SCR) of nitric oxide with ammonia6,7. In addition, the photocatalytic activity of vanadium oxide catalysts has been reported.8 Niobium oxide catalysts exhibit poor performance as redox site during oxidation reactions.9 Niobia possesses primarily (Lewis) acidic characteristics.10 However, it remarkably enhances the catalytic activity and selectivity of several metal oxide catalysts, when used as additive or support (due to the higher acidity features).11,12 Studies on the molecular structure of supported VOx and NbOx have revealed that the supported metal oxide phase forms a two-dimensional overlayer on oxide supports such as silica alumina, zirconia, or titania.4 There is a general consensus that, at low surface coverages, the vanadia species are present in an isolated tetrahedral VO4 coordination, possessing a terminal Vd O and three bridging V-O-support bonds. At higher surface * To whom correspondence should be addressed. E-mail: baltes@ uia.ua.ac.be. † Laboratory of Adsorption and Catalysis. ‡ Fortum Oil and Gas Oy. § Department of Inorganic Chemistry and Catalysis. | Centrum voor Oppervlaktechemie en Katalyse.
vanadia coverages, the vanadia species are polymerized through bridging V-O-V bonds. With increasing surface coverage square pyramidal and octahedrally coordinated (and even crystalline) vanadia species occur. Similarly, supported niobia exhibits NbO4 species at low coverage and polymerized NbO6 species at high coverage.9,13 Tantalum oxide is used as optical coating or in electronics technology as capacitor material (because of its high dielectric constant).14-16 However, the current knowledge of the physical and chemical properties of tantala is rather limited. Hydrated tantalum oxide (Ta2O5‚nH2O) is highly acidic.17 Ushikubo et al.18,19 showed that silica-supported tantalum oxide catalyzes the vapor phase Beckmann rearrangement of cyclohexanone oxime to -caprolactam. Furthermore, it is reported that SiO2Ta2O5 mixed oxides exhibit unique properties in alcohol dehydration.18,20 In this paper, we report on the structural, chemical, and catalytic properties of SiO2-, Al2O3-, and ZrO2-supported tantalum oxide catalysts, using the supported vanadium oxide case for comparison. The supported tantala and vanadia catalysts were prepared by liquid- and gas-phase modification reactions starting from tantalum ethoxide (Ta(OC2H5)5) and vanadyl acetylacetonate (VO(acac)2), respectively. The reaction of the VO(acac)2 complex with various supports has been described in detail in earlier publications.21-27 It was demonstrated that the use of transition metal complexes in the preparation of uniform and disperse supported transition metal oxides is very successful. Similarly, Ushikubo et al.19 showed that the dispersion and the catalytic performance of the supported tantalum oxide phase is much better when the synthesis is carried out using a tantalum alkoxide rather than by using aqueous impregnation.
10.1021/jp010628b CCC: $20.00 © 2001 American Chemical Society Published on Web 06/07/2001
6212 J. Phys. Chem. B, Vol. 105, No. 26, 2001 To investigate the TaOx structure and surface properties, supported vanadia-tantala mixed oxides are prepared as well. The effect of Ta addition on the nature and surface properties of supported vanadia catalysts is discussed. Chemisorption of pyridine followed by IR characterization was applied to study the acidic properties of the catalysts. The oxidation of methanol was performed in order to investigate the redox behavior. Experimental Section The silica (Grace 432), γ-alumina (AKZO), and zirconia (MEL Chemicals) supports have a specific surface area (SBET) of 300, 200, and 59 m2/g, respectively, after pretreatment in air at 500 °C for 16 h. In the liquid phase reactions, the thermally treated support was stirred at room temperature for 1 h in a zeolite-dried toluene solution containing tantalum ethoxide (Ta(OC2H5)5, Fluka) or vanadyl acetylacetonate (VO(acac)2, Fluka) for modification with Ta or V, respectively. The reaction vessel was purged with a stream of dry nitrogen. After the reaction the modified support was filtered, washed with fresh solvent and dried under vacuum at room temperature. Subsequently, the samples were calcined in air at 500 °C for 16 h. The gas-phase reactions are performed by the Atomic Layer Deposition technique (ALD).28-30 The thermally pretreated supports (3 g) were additionally preheated at the reaction temperature for 3 h in a fixed-bed flow-type reactor, under a pressure of 5-10 kPa in a nitrogen flow. After this pretreatment, the reactant vapors were passed through the support bed, with nitrogen used as carrier gas. Vaporization of the reactants was carried out by sublimation of VO(acac)2 at 160 °C and evaporation of Ta(OC2H5)5 at 200 °C, respectively. VO(acac)2 was deposited at a reaction temperature of 160 °C, Ta(OC2H5)5 at 200 °C. After the reaction, the modified support was purged with nitrogen at the reaction temperature for 2 h. Subsequently, calcination was performed in the reactor at 500 °C, in an airflow for 6h. The liquid- and gas-phase preparations of the supported vanadia-tantala mixed oxides were carried out similarly, by successive modification of the support (followed by calcination before the next modification). The Ta and V concentrations were determined by electron microprobe analysis and neutron activation analysis. In the liquid-phase modification, samples were prepared with varying metal loading (up to 0.8 mmol/g Ta and up to 1.0 mmol/g V), by using appropriate amounts of the transition metal precursors in the reaction solution. In the gas-phase modification, one reaction cycle was performed (i.e., deposition of the precursor until saturation of the support and subsequent calcination), which leads to V and Ta coverages listed in Table 1. The actual deposition process is discussed in detail in earlier publications.21-25,28-30 Pyridine adsorption was performed after calcination of the samples at 500 °C for 16h, and subsequent degassing at 150 °C for 3 h. The samples were exposed to pyridine vapor for 30 min. After evacuation for 3 h at 150 °C, the FTIR-PAS spectra were recorded. Photoacoustic infrared spectra (FTIR-PAS) were measured on a Nicolet 20 SX spectrometer, equipped with a McClelland Photoacoustic cell. X-ray diffraction patterns were collected on a Philips PW1840 powder diffractometer (45 kV, 30 mA), using Ni-filtered CuKR radiation. UV-vis diffuse reflectance spectra (UV-vis-DRS) were obtained on a UNICAM 8700 UV-vis spectrometer, with a diffuse reflectance accessory using BaSO4 standard white as reference. FT-Raman spectra were recorded
Baltes et al. TABLE 1: Ta and V Loading of the Liquid Phase and Gas Phase Prepared Supported Oxide Catalysts liquid-phase modification catalyst VOx/SiO2 VOx/Al2O3 VOx/ZrO2 TaOx/SiO2 TaOx/Al2O3 TaOx/ZrO2 VOx-TaOx/SiO2 VOx-TaOx/Al2O3 VOx-TaOx/ZrO2
V loading (mmol/g)
Ta loading (mmol/g)
up to 1.00 up to 2.00 up to 1.00
up to 1.00 up to 2.00 up to 1.00
gas-phase modification V loading (mmol/g)
Ta loading (mmol/g)
0.49 0.91 0.31 up to 0.80 up to 0.80 up to 0.60 up to 0.80 up to 0.80 up to 0.60
0.51 0.87 0.29
0.48 0.55 0.16 0.54 0.34 0.15
on a Nicolet Nexus FT-Raman spectrometer with a Ge detector. All samples were measured at room temperature in a 180 degree reflective sampling configuration, with a 1064 nm Nd:YAG excitation laser. The oxidation of methanol was carried out at 250 °C in a fixed-bed continuous flow reactor at atmospheric pressure. The amount of catalyst was controlled to obtain less than 10% methanol conversion. Prior to the reaction the catalyst was treated by flowing a He/O2 mixture at 400 °C for 16 h, to remove physisorbed water. A 100 mL/min He/O2/MeOH (88/ 8/4 mol %) flow was used as reactant mixture. Reaction products were analyzed by gas chromatography on a 2 m × 1/8 in. Porapak T column (Alltech Inc.). Results (1) Structural Characterization. (A) XRD. X-ray diffraction analysis was performed on the support materials and on the supported oxide catalysts. The silica and alumina supports are amorphous, whereas the zirconia support shows various reflections typical for the monoclinic crystal structure. The catalysts demonstrate a featureless XRD pattern, except for the characteristic pattern of the respective support materials, which suggests that the deposited oxide layers are amorphous. (B) UV-Vis Diffuse Reflectance Spectroscopy. UV-vis-DRS is used to investigate the coordination of the supported oxide species, since the energy of the oxygen-to-metal charge transfer (CT) bands is dependent on the transition metal ion and on the number of oxygen atoms surrounding it. Therefore, UV-visDRS allows to distinguish between different types of transition metal oxide configurations present on the support. Supported vanadium oxides generally show several CT bands as a function of the vanadium loading, according to the presence of three main types of configurations:25,31,32 tetrahedrally coordinated monomers and polymeric chains (characterized by CT bands between 200 and 400 nm), two-dimensional ribbons with a square pyramid coordination geometry (CT band at 410 nm) and octahedrally coordinated species (CT band at 470 nm). Figure 1 shows the UV-vis-DRS spectra of the tantalum oxide containing catalysts, prepared by liquid phase and gas phase. Pure Ta2O5 (where the central Ta5+ ion is in a octahedral configuration) shows a main absorption band at ∼270 nm. (i) SiO2. Silica-supported tantala (Figure 1A) exhibits at low loading (0.6 mmol/g, strong Raman features at 994, 697, 286, and 147 cm-1 (Figure 4c) reveal the formation of crystalline V2O5 (the XRD patterns could not evidence the presence of V2O5, which means that Raman spectroscopy is much more sensitive to small crystalline clusters). Silica-supported tantala (Figure 4d) shows a band at 974 cm-1. The infrared study also showed a distinct band around 976 cm-1. These bands are probably arising from TadO vibrations. Samples with Ta loadings of up to 0.8 mmol/g still show no bands of crystalline Ta2O5 (Raman bands26,37 at 105, 253, and 627 cm-1). (ii) ZrO2. The spectrum of the monoclinic zirconia sample has several strong bands at 634, 616, 476, 382, 334, 190, and 179 cm-1 (Figure 5a). Due to the influence of the oxide support the surface vanadium oxide band positions are shifted in comparison to the spectra of the silica-supported vanadia samples. On zirconia, the characteristic VdO band is observed at 1025 cm-1 (Figure 5b). Upon higher loadings, this band is gradually shifted to 1034 cm-1 (Figures 5c-e). In addition, broad bands appear around 830 cm-1 for low V loadings (Figure 5c), and at ∼870 and ∼930 cm-1 for higher V loadings (Figure 5, panels d and e). At low loadings, the Raman features indicate dispersed vanadia species. The bands at 1025 and 820 cm-1 are assigned to orthovanadate-like species in a tetrahedral coordination. The shift of the VdO band to higher values and the appearance of the 870 cm-1 band are due to the presence of dimeric or pyrovanadate-like (V2O7) species. At even higher coverages, the broad band around 930 cm-1 indicates the formation of polymerized surface vanadium oxide species, linked to each other by V-O-V bridges. At a V loading of 1 mmol/g (Figure 5f), bands at 994, 701, and 408 reveal the formation of V2O5 and a band at 774 shows the presence of a small ZrV2O7 phase, which is due to zirconia migration into the V2O5 crystallites.38 Remarkably, samples up to a loading of 0.7 mmol/g (spectrum 5e) exhibit no bands indicative of the formation of ZrV2O7. Many other investigations already observe a quite strong ZrV2O7 band at intermediate loadings.27,39 It demonstrates that the preparation method starting form VO(acac)2 is beneficial in the creation of highly dispersed supported vanadium oxides. The zirconia-supported tantala (Figure 5g) shows a broad Raman feature around 830 cm-1. The intensity of this band increases with higher loadings. It is attributed to the surface oxide species linked to eachother by Ta-O-Ta species. The infrared study of the supported tantala samples at quite high coverages showed the presence of an amorphous Ta2O5 phase (IR band at 2335 cm-1). The presence of microcrystalline tantalum oxide clusters, however, cannot be inferred from the Raman spectra due to the strong bands of the zirconia support. (iii) VOx-TaOx. The spectra of the supported vanadia-tantala mixed oxide catalysts (Figure 4e and Figure 5h) are similar to the supported vanadia samples. Suprisingly, they show no Raman features arising from the tantala phase. This is probably due to the low cross-section of the tantalum-oxygen vibrations compared to the vanadium-oxygen vibrations, as reported for the niobia-vanadia system.11 No changes in the Raman spectra are observed upon addition of tantalum oxide to the silicasupported vanadia samples, nor when the vanadia phase is deposited on a previously prepared silica-supported tantala phase (the VdO band is observed at 1042 cm-1). In the case of the zirconia samples (Figure 5h), the VdO band is somewhat shifted to higher values upon addition of tantala, but there is no gradual change with increasing Ta loading. In addition, the vanadiatantala samples exhibit no bands attributed to polymeric
Baltes et al. V-O-V species (at 930 cm-1), which means that the vanadia phase rests well-dispersed over the support surface and is not forced to aggregate upon Ta addition. (2) Surface Properties. (A) Acidic Properties: Adsorption of Pyridine. The spectroscopic study of adsorbed pyridine is a useful method to determine the catalysts’ surface acidity, as it enables to distinguish between Brønsted and Lewis acid sites. The type of interaction between the adsorbed pyridine and the acid sites on the catalyst surface gives rise to several characteristic infrared bands. Typical IR absorption bands of pyridine adsorbed on Brønsted acid sites (BPy) are found around 1491, 1545, 1578, and 1640 cm-1, whereas Lewis acidity gives rise to pyridine bands (LPy) around 1450, 1491, 1578, and 1612 cm-1.10,40-42 The ratio of the areas under the bands around 1545 cm-1 (BPy) and 1450 cm-1 (LPy) is usually related to the ratio of concentrations of Brønsted to Lewis acidity (B/L).41 The exact positions of the 1450 and 1612 cm-1 LPy bands are sensitive to the strength of the Lewis acidity. Figure 6 presents the infrared spectra of the supports and supported metal oxides under study, after adsorption of pyridine and degassing at 150 °C (in order to avoid the presence of physisorbed pyridine). SiO2 exposed to pyridine vapor (Figure 6A, spectrum a) shows no characteristic bands of the adsorbed base, indicating that the silica surface exhibits no Brønsted nor Lewis acidity. In contrast, alumina contains strong Lewis acid sites, giving rise to LPy bands at 1449, 1491 and 1614 cm-1 (Figure 6B, spectrum a). The infrared spectrum of the zirconia support after adsorption of pyridine (Figure 6C, spectrum a), shows bands at 1445, 1488, and 1608 cm-1, attributed to Lewis acidity. In addition, a small contribution at 1542 cm-1 indicates the presence of Brønsted sites (surface OH groups). From the position of the LPy bands it is inferred that the Lewis sites on zirconia are clearly weaker than the Lewis sites on alumina. The supported vanadium oxide shows bands typical of both Lewis and Brønsted acidity. Supported tantalum oxide catalysts exhibits only LPy bands at low coverages. At higher Ta loadings Brønsted acidity occurs as well (in accordance with the formation of Ta-OH species at higher loadings, see infrared discussion above). Still, from the Brønsted-to-Lewis ratio (B/ L) obtained after integration of the bands around 1545 cm-1 (BPy) and 1450 cm-1 (Lpy) (presented in Table 2), it can be inferred that the supported tantala samples exhibit more Lewis acidity in comparison to supported vanadia. In addition, the supported vanadia-tantala mixed oxides show an increased Lewis acidity (B/L decreases), with increasing Ta content. Clearly, supported tantalum oxide exhibits predominantly Lewis acidic characteristics. Furthermore, the position of the LPy bands indicate, that the Lewis acid sites on the surface of the silica and alumina-supported tantala catalysts are stronger than the Lewis acid sites on TaOx/ZrO2. (B) Catalytic Performance in the Oxidation of Methanol. The oxidation of methanol is a useful test reaction to study the redox behavior of the supported oxide catalysts. It is commonly used as a catalytic surface probe, as it leads to various products according to the chemical composition or arrangement of the surface atoms.4,43,44 The product selectivity reflects the presence of acidic, basic, or redox sites. Acidic sites give rise to dimethyl ether, basic sites produce carbon oxides (COx) and redox sites produce formaldehyde, methyl formate, and dimethoxymethane. In Tables 2, 3, and 4 the results are presented of the oxidation of methanol for the catalysts under study. (i) SiO2. Pure silica shows very poor catalytic performance. The activity of the silica-supported vanadium oxide is signifi-
Tantalum Oxide and Vanadia-tantala Mixed Oxides
J. Phys. Chem. B, Vol. 105, No. 26, 2001 6217
Figure 6. FTIR spectra of pyridine adsorbed on (A) silica-supported oxide catalysts; (a) pure SiO2; (b) TaOx/SiO2 (0.21 mmol/g); (c) TaOx/SiO2 (0.47 mmol/g); (d) VOx/SiO2 (0.37 mmol/g); (e) VOx-TaOx/SiO2 with respective V and Ta loading of 0.37 and 0.20 mmol/g, and (B) aluminasupported oxide catalysts; (a) pure Al2O3; (b) TaOx/Al2O3 (0.20 mmol/g); (c) TaOx/Al2O3 (0.80 mmol/g); (d) VOx/Al2O3 (0.70 mmol/g); (e) VOxTaOx/Al2O3 with respective V and Ta loading of 0.72 and 0.45 mmol/g. C: Zirconia-supported oxide catalysts; (a) pure ZrO2; (b) TaOx/ZrO2 (0.24 mmol/g); (c) TaOx/ZrO2 (0.60 mmol/g); (d) VOx/ZrO2 (0.37 mmol/g); (e) VOx-TaOx/ZrO2 with respective V and Ta loading of 0.37 and 0.34 mmol/g.
TABLE 2: Activity and Selectivity of the SiO2 Supported VOx, TaOx and VOx-TaOx Catalysts catalyst
V (mmol/g)
Ta (mmol/g)
SiO2 VOx/SiO2 VOx/SiO2 VOx/SiO2 VOx/SiO2 TaOx/SiO2 TaOx/SiO2 VOx-TaOx/SiO2f VOx-TaOx/SiO2f VOx-TaOx/SiO2f TaOx-VOx/SiO2g
0.00 0.20 0.36 0.70 1.03 0.00 0.00 0.36 0.35 0.24 0.37
0.00 0.00 0.00 0.00 0.00 0.24 0.47 0.20 0.59 0.60 0.20
(10-3
Aa mol/g h)
TOFb (10-3 s-1)
1.00 4.99 5.23 5.98 5.23 3.99 5.28 5.48 6.03 5.08 5.98
6.92 4.04 2.37 1.41 4.62 3.12
TOFredoxc (10-3 s-1) 6.72 3.83 1.80 0.89 0.00 0.94
TOFacidd (10-3 s-1) 0.20 0.21 0.53 0.35 4.62 2.19
selectivity (%)e COx
FA
DME
MF
0.0 0.0 0.0 1.9 12.3 0.0 0.0 0.0 0.0 0.0 0.0
2.3 25.1 45.2 47.1 45.8 0.0 14.0 74.8 54.8 47.5 74.3
86.1 2.9 5.1 22.3 24.6 100.0 70.0 25.2 40.3 26.6 25.0
11.6 72.0 49.7 28.7 17.3 0.0 16.0 0.0 4.9 25.9 0.7
a Activity, moles of methanol converted per gram of catalyst and per hour. b Turnover frequency, moles of methanol converted per mole of metal and per second. c Turnover frequency based on the amount of redox products (FA + MF). d Turnover frequency based on the amount of DME.e FA, formaldehyde; DME, dimethyl ether; MF, methyl formate. f Sample prepared by deposition of TaOx prior to deposition of VOx. g Sample prepared by deposition of VOx prior to deposition of TaOx.
cantly higher. VOx/SiO2 produces mainly oxidation products (formaldehyde and methyl formate). The high selectivity for methyl formate substantiates the high dispersion of the vanadium oxide species on the silica support.45 The catalyst surface exhibits an even distribution of vanadia and residual hydroxyl groups of the silica support. Methanol is adsorbed and transformed to formaldehyde on the vanadium sites. However, methanol can react with the residual surface hydroxyl sites as well, which results in adsorbed methoxy groups.46,47 Therefore, during the methanol conversion, the formaldehyde formed on the vanadia sites is in close interaction with methoxy groups on the silica sites. Via a hemiacetal intermediate, methyl formate is produced.45 With increasing loadings, the formaldehyde selectivity increases, as the surface density of the residual SiOH groups decreases. For the silica samples with V loadings at which crystalline V2O5 is observed (>0.6 mmol/g), the
formation of redox products markedly decreases, in favor of the formation of dimethyl ether and carbon oxides. Silica-supported tantalum oxide catalysts show activities comparable to the VOx/SiO2. However, the main reaction product is dimethyl ether, indicating the acidic properties of the TaOx phase. Remarkably, at higher Ta loading the TaOx/ SiO2 samples produce methyl formate and formaldehyde along with the formation of dimethyl ether. This suggests that the interaction of tantala with the silica support gives rise to a considerable redox activity too. (ii) Al2O3. Alumina is very active for the dehydration of methanol to dimethyl ether, due to the strong acidity of the alumina surface sites. Deposition of vanadium oxide leads to a decreasing activity and increasing formaldehyde selectivity as a function of the vanadium loading. Moreover, the main product of the reaction is still dimethyl ether, which suggests that the
6218 J. Phys. Chem. B, Vol. 105, No. 26, 2001
Baltes et al.
TABLE 3: Activity and Selectivity of the Al2O3 Supported VOx, TaOx, and VOx-TaOx Catalysts catalyst
V (mmol/g)
Ta (mmol/g)
Aa (mol/g‚h)
Al2O3 VOx/Al2O3 VOx/Al2O3 VOx/Al2O3 VOx/Al2O3 TaOx/Al2O3 TaOx/Al2O3 VOx-TaOx/Al2O3f TaOx-VOx/Al2O3g
0.00 0.25 0.70 1.20 2.10 0.00 0.00 0.72 0.69
0.00 0.00 0.00 0.00 0.00 0.20 0.80 0.45 0.41
0.96 0.97 0.85 0.70 0.50 0.78 0.53 0.50 0.50
TOFb (s-1)
TOFredoxc (s-1)
TOFacidd (s-1)
1.07 0.34 0.16 0.07 1.08 0.18
0.02 0.05 0.04 0.02 0.00 0.00
1.05 0.29 0.12 0.04 1.08 0.18
selectivity (%)e COx
FA
DME
MF
0.0 0.0 0.0 0.0 2.0 0.0 0.0 0.0 0.0
0.0 2.3 14.0 21.7 29.1 0.0 0.0 13.1 12.8
100.0 97.7 86.0 77.3 67.9 100.0 100.0 86.9 87.2
0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0
a Activity, moles of methanol converted per gram catalyst and per hour. b Turnover frequency, moles of methanol converted per mole of metal and per second. c Turnover frequency based on the amount of redox products (FA + MF). d Turnover frequency based on the amount of DME.e FA, formaldehyde; DME, dimethyl ether; MF, methyl formate. f Sample prepared by deposition of TaOx prior to deposition of VOx. g Sample prepared by deposition of VOx prior to deposition of TaOx.
TABLE 4: Activity and Selectivity of the ZrO2 Supported VOx, TaOx, and VOx-TaOx Catalysts catalyst
V (mmol/g)
Ta (mmol/g)
Aa (mol/g‚h)
ZrO2 VOx/ZrO2 VOx/ZrO2 VOx/ZrO2 VOx/ZrO2 TaOx/ZrO2 TaOx/ZrO2 VOx-TaOx/ZrO2f VOx-TaOx/ZrO2f VOx-TaOx/ZrO2f VOx-TaOx/ZrO2f TaOx-VOx/ZrO2g
0.00 0.09 0.19 0.36 1.02 0.00 0.00 0.36 0.36 0.36 0.36 0.35
0.00 0.00 0.00 0.00 0.00 0.15 0.24 0.05 0.17 0.34 0.58 0.20
0.05 0.55 0.65 1.00 0.55 0.03 0.03 0.98 0.98 0.97 0.64 0.98
TOFb (s-1)
TOFredoxc (s-1)
TOFacidd (s-1)
1.69 0.95 0.77 0.15 0.06 0.04
1.53 0.91 0.76 0.12 0.05 0.01
0.16 0.04 0.01 0.01 0.02 0.03
selectivity (%)e COx
FA
DME
MF
0.0 0.0 0.0 0.0 10.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 68.3 87.1 92.0 74.9 0.0 0.0 92.0 92.2 91.0 88.7 92.5
15.2 9.7 3.8 1.7 8.1 28.4 83.8 4.0 4.9 6.0 7.7 5.8
84.8 22.0 9.1 6.3 6.7 71.6 16.2 4.0 2.9 3.0 3.6 1.7
a Activity, moles of methanol converted per gram of catalyst and per hour. b Turnover frequency, moles of methanol converted per mole of metal and per second. c Turnover frequency based on the amount of redox products (FA + MF). d Turnover frequency based on the amount of DME.e FA, formaldehyde; DME, dimethyl ether; MF, methyl formate. f Sample prepared by deposition of TaOx prior to deposition of VOx. g Sample prepared by deposition of VOx prior to deposition of TaOx.
alumina acid sites are much more active than the redox activity of the vanadia sites. Alumina-supported tantala shows exclusive dehydration activity. Increasing the Ta loading on alumina also results in decreasing activity, which indicates that the surface TaOx sites (although acidic) are less active than the active sites of the alumina surface. Therefore, the overall activity of the TaOx/ Al2O3 samples is attributed to the reactivity of the support sites rather than on the reactivity of the tantala sites. (iii) ZrO2. Pure zirconia exhibits some redox activity and produces methyl formate, besides some dimethyl ether rising from the acidic surface sites. Deposition of vanadium oxide significantly increases the activity and leads to a quite selective formation of formaldehyde (except for samples containing crystalline fractions of V2O5). Deposition of TaOx on zirconia leads to a slightly increased activity in comparison to the pure support. The acidic properties of the tantala sites result in a marked increase of the dimethyl ether selectivity. It is well established that supported vanadium oxide catalysts show quite different activity and selectivity when the support is changed. A readily reducible oxide support surface gives rise to a more active supported vanadium oxide phase.4,48 This means that the reactivity depends on the reduction characteristics of the V-O-support bonds rather than the characteristics of the terminal VdO bonds. As a result, the redox activity of the supported vanadium oxides increases in the order SiO2 < Al2O3 < ZrO2 = TiO2 according to the better surface reducibility of the support. This can also be inferred from the TOFredox values of the VOx/SiO2, VOx/Al2O3 and VOx/ZrO2 catalysts in Tables
TABLE 5: Summary of the Spectral Features of the Supported Vanadia and Supported Tantala Catalysts UV-VIS-DR charge transfer band assignments CT band position (nm) nature of species
VOx
TaOx
tetrahedral monomer 250, 300 tetrahedral chains 350 square pyramidal 410 octahedral 470 FTIR band assignments
230 265
IR band position (cm-1) infrared vibration
SiO2
V-OH VdO V-O-support V-O-V Ta-OH TadO Ta-O-support Ta-O-Ta
3660
Al2O3
ZrO2
3660 1025
3660 1024
3743 980, 2336
840 3743 980, 2336
930 3680 976 918
840 Raman band assignments Raman shift (cm-1)
Raman vibration
SiO2
Al2O3
ZrO2
VdO V-O-V TadO Ta-O-Ta
1042
1026
974
980
1025-1030 850-820 980 840-820
2, 3, and 4. However, the correlation between activity of the surface tantala species and the surface reducibility of the support
Tantalum Oxide and Vanadia-tantala Mixed Oxides is not like in the case of vanadia. The total activity increases in the order SiO2 < Al2O3 < ZrO2, similar to the vanadia case. However, the main product of the reaction is dimethyl ether (except for some methyl formate on the ZrO2 supported samples, which is due to the contribution of exposed zirconia redox sites), which shows that the tantalum oxide exhibits predominantly acidic characteristics. In addition, the total activity of the silicasupported tantala is almost the same as silica-supported vanadia, while the total activity of the zirconia-supported tantala is much lower than the vanadia analogue. Clearly, the reducibility of the support surface has a much stronger effect on the surface vanadia species than on the surface tantala sites. (iV) VOx-TaOx. The difference in reactivity between vanadia and tantala is also reflected in the supported mixed oxide catalysts. The addition of tantala to silica-supported vanadia increases the total activity and the selectivities toward formaldehyde and dimethyl ether. On the other hand, the addition of tantala to alumina- and zirconia-supported vanadia does not affect the catalytic performance, which substantiates that on readily reducible supports the vanadia sites are much more active than the tantala sites. In addition, reversing the sequence of deposition of VOx and TaOx does not lead to significant changes in activity and selectivity. The only exception is the zirconiasupported catalyst prepared by the deposition of vanadium oxide on the amorphous tantalum overlayers (samples with high Ta loadings). The activity of this sample is significantly lower, although the selectivity pattern is the same as for the other zirconia-supported mixed vanadia-tantala. In this case a part of the vanadia sites is linked to the tantalum oxide overlayer. Apparently, the poor reducibility of the tantala phase leads to a decreased activity of these vanadia sites. Discussion Molecular Structures and Surface Properties of Tantalum Oxide Containing Catalysts. The supported tantalum oxide catalysts investigated clearly show the existence of two types of surface configurations, depending on the properties of the oxide support and the surface coverage. The spectral features of the supported oxide catalysts obtained in this study are summarized in Table 5. TaOx/SiO2. The spectroscopic study of the silica-supported tantala catalyst shows at low Ta loading a highly uniform surface structure, characterized by a charge-transfer band in the UVvis-DRS spectra at 233 nm, and TadO bands at 974 and 976 cm-1 in Raman and infrared, respectively. The TadO species are linked to the silica support by Si-O-Ta bridges, which is inferred from the consumption of Si-OH groups after deposition of the tantala phase, and the appearance of a broad IR band at 918 cm-1. A recent EXAFS study of silica-supported tantalum oxide49 revealed that the surface oxide species are present as isolated tantalate species with one short Ta-O bond and three long Ta-O bonds. This means that at low loadings the deposited TaOx species are in an isolated tetrahedral configuration. However, the spectroscopic results from this study suggest that at higher loadings an additional surface configuration occurs, which gives rise to a CT band around 260 nm and exhibits TaOH species (IR band at 3680 cm-1). On the other hand, the octahedrally coordinated Ta in pure Ta2O5 shows similar spectroscopic features. Therefore, the second surface structure is attributed to octahedrally coordinated tantala, which is probably polymerized via Ta-O-Ta bridges. Furthermore, it seems that higher Ta loadings can be achieved by the ALD gas phase technique in comparison to the liquid phase synthesis, before octahedral (and probably polymerized) tantala species occur. The difference between gas and liquid
J. Phys. Chem. B, Vol. 105, No. 26, 2001 6219 phase can be explained by the structure of the tantalum ethoxide precursor. Most simple tantalum alkoxides adopt binuclear structures (Ta2(OR)8(µ-OR)2), with two bridging alkoxide ligands. Evaporation of the alkoxide occurs only after dissociation to monomers (Ta(OR)5).16 This means that during the ALD process monomeric species are interacting with the support surface, resulting in a much better distribution of isolated tantalum centers over the support surface. Still, liquid-phase samples with loadings up to 0.8 mmol/g (15 wt % Ta) show no evidence for the formation of crystalline Ta2O5. The pyridine adsorption experiments showed that the silicasupported tantala phase is mainly Lewis acidic, although at high loadings also Brønsted acidity occurs, due to the presence of the Ta-OH groups. In the oxidation of methanol the supported tantala catalysts show excellent dehydration performance due to the acidic properties, but at higher loadings the interaction between tantala and silica causes some redox activity as well. This may be due to the ease of hydrolysis of the bridging TaO-Si bonds during the methanol oxidation, as reported for the silica-supported niobia case.46,50 However, a detailed understanding for this redox behavior is not yet established and a thorough oxidation study is required in order to clarify this issue. TaOx/Al2O3. The alumina-supported tantalum oxides also exhibit CT band around 235 nm which indicates that the tantalum species are present in an isolated tetrahedral configuration. However, absorption at 265 nm suggests that even at low coverages a minor part of the TaOx is octahedrally coordinated, even in the case of the ALD-prepared samples. TadO species are observed by an IR absorption at 980 cm-1. At higher coverages an amorphous tantalum oxide overlayer exists, giving rise to an IR band at 2336 cm-1 and Ta-OH species at 3743 cm-1 (the band at 3680 cm-1 which is also associated with Ta-OH cannot be distinguished unambiguously due to the IR bands of the support hydroxyls). The Ta sites supported on alumina show Lewis acidity and an exclusive dehydration activity in the oxidation of methanol. However, the alumina sites exhibit much higher activity than the tantala sites. TaOx/ZrO2. On zirconia, the majority of the tantala species is in a polymeric octahedral coordination, even at low loadings, which is substantiated by the CT band at 270 nm and the TaO-Ta vibrations around 820-830 cm-1 in infrared and Raman. Similar to the alumina samples, the zirconia-supported tantala shows IR absorption at 2335 and 3743 cm-1 at higher coverages, indicating the presence of an amorphous Ta2O5 layer. Zirconia-supported tantalum oxide is also Lewis acidic, which results in an almost exclusive dehydration activity in the oxidation of methanol. It is concluded that at low Ta coverage the tantala species have exclusively Td symmetry on silica, whereas both Td and Oh symmetry of Ta is observed on alumina, and mainly Oh symmetry of Ta is observed on zirconia. This can be explained by the fact that metal ions in an overlayer at low coverage assume the symmetry of the support metal ions. It is known that silica shows only Td symmetry,51 whereas alumina shows a surface constructed of both Td and Oh symmetries,52,53 and zirconia shows only Oh symmetry54 if one considers both nearest and next-nearest oxygen atoms. Supported VOx-TaOx Mixed Oxides. The spectroscopic and catalytic comparison between supported vanadia and tantala oxide was very useful to deduce valuable information on the tantalum oxide structure and properties. The spectroscopic study of the supported mixed oxides leads to similar conclusions as in the case of the supported tantala and supported vanadia as
6220 J. Phys. Chem. B, Vol. 105, No. 26, 2001 such. Interestingly, the structure and dispersion of one metal oxide species is nearly unaffected by the presence of the other metal oxide species nor by the grafting sequence of the oxides. The catalytic properties are significantly determined by the oxide support. Silica-supported vanadia-tantala catalysts exhibit high selectivity toward formaldehyde and dimethyl ether, according to the presence of both active redox and active acid centers. In contrast, on a readily reducible support like ZrO2, the vanadia species are much more active than the tantala sites. In fact, the oxidation of methanol requires only one active surface site; therefore, the catalytic performance of the zirconiasupported catalysts is not promoted by the tantala species. Consequently, a mixed vanadia-tantala catalyst is useful for reactions that require both active redox and efficient acidic surface sites. Conclusions (1) Supported tantalum oxide show basically two types of molecular structures. On silica at low coverages, isolated tetrahedral species occur, with a short TadO bond and three bridging Ta-O-support bonds. At higher coverages, octahedrally coordinated tantala species occur. The alumina-supported tantala shows both surface species already at low coverages, whereas the zirconia samples predominantly exhibit polymerized, octahedrally configured tantalum oxide. In addition, the formation of an amorphous surface overlayer is revealed on both alumina and zirconia at high tantala coverages. (2) The atomic layer deposition of the tantalum ethoxide precursor results in a better uniformity of the surface tantala species in comparison to the liquid phase. This is due to the binuclear structure of the tantalum alkoxide in the liquid phase, and the dissociation to a monomeric structure after evaporation. (3) The supported tantala catalysts possess unique surface properties. The infrared study of the adsorption of pyridine showed the Lewis acidic characteristics of the surface tantala species. This is also reflected in the oxidation of methanol, which reveals that the supported tantala shows unique acidic properties. Suprisingly, the interaction of tantala with the silica support induces redox properties too. (4) Supported vanadia-tantala mixed oxides are very well dispersed on the support surfaces. The grafting sequence does not lead to significant changes in surface structure or methanol oxidation performance. The ternary vanadia-tantala-support catalyst system is potentially useful in reactions requiring both redox and acid sites. Acknowledgment. This work was supported by the FWO Vlaanderen (Research project G.0446.99). M. Baltes thanks the European Science Foundation (ESF) for a short-term fellowship in the scope of the ESF network on “Elementary Steps of Layer Growth in the Fabrication of NoVel Materials by Atomic Layer Epitaxy”, for a research stay at Fortum Oil and Gas Oy, Espoo, Finland. P. Van Der Voort is a postdoctoral fellow of the F.W.O. Vlaanderen. The authors are grateful to Mrs. Minna Helsto and Mrs. Mirja Rissanen for their aid in the experimental work. References and Notes (1) Koranne, M.; Goodwin, J. G.; Marcelin, G. J. Catal. 1994, 148, 388. (2) Dias, C. R.; Portela, M. F.; Bond, G. C. J. Catal. 1995, 157, 344. (3) Bond, G. C.; Tahir, S. F. Appl. Catal. 1991, 71, 1. (4) Deo, G.; Wachs, I. E.; Haber J. Crit. ReV. Surf. Chem. 1994, 4, 141. (5) Dunn, J. P.; Stenger, H. P.; Wachs, I. E. Catal. Today 1999, 51, 301. (6) Vogt, E. T. V.; Boot, A.; van Dillen, A. J.; Geus, J. W.; Janssen, F. J. J. G.; van den Kerkhof, F. M. G. J. Catal. 1988, 114, 313.
Baltes et al. (7) Wachs, I. E.; Deo, G.; Weckhuysen, B. M.; Andreini, A.; Vuurman, M. A.; de Boer, M.; Amiridis, M. D. J. Catal. 1996, 161, 211. (8) Tran, K.; Hanning-Lee, M. A.; Biswas, A.; Stiegman, A. E.; Scott, G. W. J. Am. Chem. Soc. 1995, 117, 2618. (9) Wachs, I. E.; Briand, L. E.; Jehng, J. M.; Burcham, L.; Gao, X. Catal. Today 2000, 57, 323. (10) Datka, J.; Turek, A. M.; Jehng, J. M.; Wachs, I. E. J. Catal. 1992, 135, 186. (11) Deo, G.; Wachs, I. E. J. Catal. 1994, 146, 335. (12) Viparelli, P.; Ciambelli, P.; Lisi, L.; Ruoppolo, G.; Russo, G.; Volta, J. C. Appl. Catal., A 1999, 184, 291. (13) Jehng, J. M.; Wachs, I. E. J. Phys. Chem. 1991, 95, 7373. (14) Kukli, K.; Ritala, M.; Matero, R.; Leskela¨, M. J. Cryst. Growth 2000, 212, 459. (15) Meng, J. F.; Rai, B. K.; Katiyar, R. S.; Bhalla, A. S. J. Phys. Chem. Solids 1997, 58, 1503. (16) Pollard, K. D.; Puddephatt, R. J. Chem. Mater. 1999, 11, 1069. (17) Ushikubo, T. Catal. Today 2000, 57, 331. (18) Ushikubo, T.; Wada, K. J. Catal. 1994, 148, 138. (19) Ushikubo, T.; Wada, K. Appl. Catal., A 1995, 124, 19. (20) Guiu, G.; Grange, P. J. Catal. 1995, 156, 132. (21) Van Der Voort, P.; Babitch, I. B.; Grobet, P. J.; Verberckmoes, A. A.; Vansant, E. F. J. Chem. Soc., Faraday Trans. 1996, 92, 3635. (22) Baltes, M.; Van Der Voort, P.; Collart, O.; Vansant, E. F. J. Porous Mater. 1998, 5, 357. (23) Baltes, M.; Collart, O.; Van Der Voort, P.; Vansant, E. F. Langmuir 1999, 15, 5841. (24) Baltes, M.; Van Der Voort, P.; Weckhuysen, B. M.; Rao, R. R.; Catana, G.; Schoonheydt, R. A.; Vansant, E. F. Phys. Chem. Chem. Phys. 2000, 2, 2673. (25) Baltes, M.; Cassiers, K.; Van Der Voort, P.; Weckhuysen, B. M.; Schoonheydt, R. A.; Vansant, E. F. J. Catal. 2001, 197, 160. (26) Huuhtanen, J.; Sanati, M.; Andersson, A.; Andersson, S. L. T. Appl. Catal A 1993, 97, 197. (27) Prinetto, F.; Ghiotto, G.; Occhiuzzi, M.; Indovina, V. J. Phys. Chem. B 1998, 102, 10316. (28) Haukka, S.; Kyto¨kivi, A.; Lakomaa, E.-L.; Lehtovirta, U.; Lindblad, M.; Lujala, V.; Suntola, T. In Preparation of Catalysts VI, Scientific Bases for the Preparation of Heterogeneous Catalysts; Poncelet, G., Ed.; p 957, Elsevier Science, Amsterdam, 1995; p 957. (29) Kyto¨kivi, A.; Lakomaa, E.-L.; Root, A.; O ¨ sterholm, H.; Jacobs, J.-P.; Brongersma, H. H. Langmuir 1997, 13, 2717. (30) Kyto¨kivi, A.; Jacobs, J.-P.; Hakuli, A.; Merila¨inen, J.; Brongersma, H. H. J. Catal. 1996, 162, 190. (31) Schraml-Marth, M.; Wokaun, A.; Pohl, M.; Krauss, H. L. J. Chem. Soc., Faraday Trans 1991, 87, 2635. (32) Scharf, U.; Schraml-Marth, M.; Wokaun, A.; Baiker, A. J. Chem. Soc., Faraday Trans. 1991, 87, 3299. (33) Zhang, J.-Y.; Lim, B.; Boyd, I. W. Thin Solid Films 1998, 336, 340. (34) Vuurman, M. A.; Stufkens, D. J.; Oskam, A.; Deo, G.; Wachs, I. E. J. Chem. Soc., Faraday Trans. 1996, 92, 3259. (35) Ono, H.; Koyanagi, K. Appl. Phys. Lett. 2000, 77, 1431. (36) Kantcheva, M. Phys. Chem. Chem. Phys. 2000, 2, 3043. (37) Szanics, J.; Okubo, T.; Kakihana, M. J. Alloy Compd. 1998, 281, 206. (38) Su, S. C.; Bell, A. T. J. Phys. Chem. 1998, 102, 7000. (39) Sanati, M.; Andersson, A.; Wallenberg, L. R.; Rebenstorf, B. Appl. Catal., A 1993, 106, 51. (40) Kataoka, T.; Dumesic, J. A. J. Catal. 1988, 112, 66. (41) Rajagopal, S.; Marzari, J. A.; Miranda, R. J. Catal. 1995, 151, 192. (42) Damyanova, S.; Grange, P.; Delmon, B. J. Catal. 1997, 168, 421. (43) Tatiboue¨t, J. M. Appl. Catal., A 1997, 148, 213. (44) Forzatti, P.; Tronconi, E.; Elmi, A. S.; Busca, G. Appl. Catal., A 1997, 157, 387. (45) Louis, C.; Tatiboue¨t, J. M.; Che, M. J. Catal. 1988, 109, 354. (46) Jehng, J.-M.; Hu, H.; Gao, X.; Wachs, I. E. Catal. Today 1996, 28, 335. (47) Gao, X.; Bare, S. R.; Weckhuysen, B. M.; Wachs, I. E. J. Phys. Chem. B 1998, 102, 10842. (48) Deo, G.; Wachs, I. E. J. Catal. 1994, 146, 323. (49) Tanaka, T.; Nojima, H.; Yamamoto, T.; Takenaka, S.; Funabiki, T.; Yoshida, S. Phys. Chem. Chem. Phys. 1999, 1, 5235. (50) Wachs, I. E.; Jehng, J.-M.; Deo, G.; Hu, H.; Arora, N. Catal. Today 1996, 28, 199. (51) Vansant, E. F.; Van Der Voort, P.; Vrancken, K. C. Characterization and Chemical Modification of the Silica Surface, Studies in Surface Science and Catalysis; Elsevier: Amsterdam, 1995; Vol. 93. (52) Peri, J. B. J. Phys. Chem. 1966, 70, 3168. (53) Kno¨zinger, H.; Ratnasamy, P. Catal. ReV. Sci. Eng. 1978, 17, 31. (54) Clearfield, A.; Serette, G. P. D.; Khazi-Syed, A. H. Catal. Today 1994, 20, 295.
